检索结果-内蒙古大学图书馆

VFL+: Low-Coupling Vertical Federated Learning With Privileged Information Paradigm

IEEE Transactions on Emerging Topics in Computational Intelligence 2025年

作者： Dai, Wei Cui, Teng Zhang, Tong Chen, Badong China University of Mining Technology School of Information and Control Engineering Xuzhou221116 China South China University of Technology School of Computer Science and Engineering Guangzhou510006 China Xi'an Jiaotong University Institute of Artificial Intelligence and Robotics Xi'an710049 China

Vertical Federated Learning (VFL) enables the construction of models by combining clients with different features without compromising privacy. Existing VFL methods exhibit tightly coupled participant parameters, resulting in substantial interdependencies among clients during the prediction phase, which significantly hampers the model's usability. To tackle these challenges, this paper studies a VFL approach with low coupling of parameters between clients. Drawing inspiration from federated cooperation and teacher-supervised learning, we propose a low-coupling vertical federated learning with privileged information paradigm (VFL+), allowing participants to make autonomous predictions. Specifically, VFL+ treats information from other clients as privileged data during the training phase rather than the testing phase, thereby achieving independence in individual model predictions. Subsequently, this paper further investigates three typical scenarios of vertical cooperation and designs corresponding cooperative frameworks. Systematic experiments on real data sets demonstrate the effectiveness of the proposed method. © 2017 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Power Consumption Optimization of GPU Server with Offline Reinforcement Learning

引用

IEEE Access 2025年 13卷 85826-85837页

作者： Chung, Heechan Im, Yeeun Park, Jongchan Lee, Taeho Kim, TaeYoung Kim, Hyungjun Lee, Dongwhan KT Research and Development Center AI Security Research Team Future Network Laboratory Network Group Seoul Korea Republic of KT Research and Development Center AI Perceived Quality Research Team Future Network Laboratory Network Group Seoul Korea Republic of Korea Advanced Institute of Science and Technology Robotics Program Daejeon Korea Republic of Korea Advanced Institute of Science and Technology School of Electrical Engineering Daejeon Korea Republic of

The rapid growth of artificial intelligence (AI) has led to increased reliance on power-intensive Graphics Processing Units (GPUs), which are essential for training and deploying large-scale models. However, the escalating energy demands of AI workloads pose sustainability challenges, necessitating efficient power management strategies to reduce carbon footprints. Optimizing GPU server power consumption is complex due to the interdependence of various components. Conventional methods often involve trade-offs: increasing fan speed enhances cooling but raises overall power usage, whereas lowering GPU clock frequencies conserves energy at the cost of longer computation times. To address these challenges, we propose a data-driven optimization framework based on offline reinforcement learning (RL). Our approach collects operational data from a custom-designed workload that simulates varying server loads, capturing key metrics such as power consumption, temperature, and core frequency. The reward function balances power efficiency with performance. The reinforcement learning agent learns from pre-collected server logs, enabling intelligent real-time GPU clock control decisions without costly live experiments. Additionally, periodic fan speed adjustments and pre-training of the Q-network further enhance overall efficiency. Experimental results demonstrate that our method reduces power consumption by 3.62% while improving computation time by 1.51% for synthetic workloads. For LLaMA-2 fine-tuning, power consumption decreases by 6.40% with only a minor 1.27% increase in computation time, demonstrating its practical effectiveness. Our framework was validated on the latest NVIDIA L40S GPU, demonstrating its compatibility with cutting-edge hardware. © 2013 IEEE.

关键词： Energy utilization

来源：评论

学校读者我要写书评

暂无评论

Constructing Semantical Structure by Segmentation Integrated Video Embedding for Temporal Action Detection

引用

IEEE Transactions on Circuits and Systems for Video Technology 2025年

作者： Zhao, Zixuan Liu, Shuming Zhao, Chengze Zhao, Xu Shanghai Jiao Tong University School of Electronic Information and Electrical Engineering Shanghai200240 China King Abdullah University of Science and Technology Computer Science Program Thuwal23599 Saudi Arabia

Video embedding is the pivot in Temporal Action Detection (TAD). Once the video embedding can robustly capture the essence of actions and perceive activities in complex scenes, the TAD model can more accurately localize action boundaries. Currently, video embedding is typically based on rule-based pixel convolution or cube-based transformer, wherein structured semantic information is intertwined, leading to the submergence of crucial spatial semantic information, such as the intrinsic motion of key semantic objects and interactions among semantic objects. To address these limitations, it is imperative to explore alternative approaches. With the remarkable performance of general semantic segmentation models in visual representation, we introduce the general segmentation model SEEM into the video embedding paradigm, constructing a semantically structured representation from perceptual semantics to cognitive semantics. To more effectively utilize SEEM for structured video representation, we designed the Semantic Adapter (Sem-Adapter) as a bridge to connect the two models. Firstly, we design a Self-Motion Module (SMM) to pay attention to the self-motion of key semantic regions. Secondly, we propose a Mutual Relation Module (MRM) to construct the interactions between semantic regions. Extensive experiments on ActivityNet-1.3, THUMOS-14 and EPIC-Kitchens-100 reveal that our method significantly outperforms state-of-the-art methods under the same input modality, and our method improves the average mAP from 60.6% to 64.2% on THUMOS-14 with the same backbone. The code is available on https://***/shouxiaozixuan/semtad. © 1991-2012 IEEE.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

SELM: From Efficient Autonomous Exploration to Long-term Monitoring in Semantic Level

引用

IEEE Transactions on Cognitive and Developmental Systems 2025年

作者： Lang, Fang Qin, Yongsen Wang, Yinchuan Liu, Jin Wang, Chaoqun Song, Wei Zhu, Qiuguo Song, Rui Zhejiang University Polytechnic Institute Hangzhou310015 China Shandong University School of Control Science and Engineering Jinan China Zhejiang University Robotics Institute Hangzhou China Zhejiang University College of Computer Science and Technology Hangzhou China

Maintaining up-to-date environmental models from initial deployment through long-term autonomy in service is critical for applications like navigation and task planning. To address the challenges of persistent monitoring in unknown environments, we introduce a Two-Stage monitoring strategy, termed the Semantic-level autonomous Exploration and Long-term environment Monitoring (SELM) framework. In the first stage, we introduce a novel semantic exploration method to adapt to new environments quickly. Leveraging the semantic information within the incrementally constructed 3D Scene Graph (3DSG), we combine the Next-Best-View (NBV) selection with room semantics, introducing a more efficient and comprehensive approach for multi-room indoor environment exploration. In addition, the exploration provides patrol routes, the room distance-connectivity graph, and complete environment initial states for subsequential monitoring. The monitoring stage aims to persistently patrol to update the world model in the presence of dynamic changes, including changes in objects' positions. We formulate the long-term monitoring problem as the Partially Observable Markov Decision Process (POMDP) to cope with the environmental uncertainty. To solve the POMDP, we propose the Graph Attention Bidirectional long short-term memory Proximal Policy Optimization (GABPPO) algorithm for the optimal patrol strategy. The feasibility and effectiveness of the proposed SELM framework are verified through extensive experiments. © 2016 IEEE.

关键词： Environmental monitoring

来源：评论

学校读者我要写书评

暂无评论

DLPO : Towards a Robust, Efficient, and Generalizable Prompt Optimization Framework from a Deep-Learning Perspective

arXiv

引用

arXiv 2025年

作者： Peng, Dengyun Zhou, Yuhang Chen, Qiguang Liu, Jinhao Chen, Jingjing Qin, Libo Research Center for Social Computing and Interactive Robotics Harbin Institute of Technology China School of Computer Science and Engineering Central South University China Fudan University China

Large Language Models (LLMs) have achieved remarkable success across diverse tasks, largely driven by well-designed prompts. However, crafting and selecting such prompts often requires considerable human effort, significantly limiting its scalability. To mitigate this, recent studies have explored automated prompt optimization as a promising solution. Despite these efforts, existing methods still face critical challenges in robustness, efficiency, and generalization. To systematically address these challenges, we first conduct an empirical analysis to identify the limitations of current reflection-based prompt optimization paradigm. Building on these insights, we propose 7 innovative approaches inspired by traditional deep learning paradigms for prompt optimization (DLPO), seamlessly integrating these concepts into text-based gradient optimization. Through these advancements, we progressively tackle the aforementioned challenges and validate our methods through extensive experimentation. We hope our study not only provides valuable guidance for future research but also offers a comprehensive understanding of the challenges and potential solutions in prompt optimization. Our code is available at https://***/sfasfaffa/DLPO. Copyright © 2025, The Authors. All rights reserved.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

An Adversarial Approach to Register Extreme Resolution Tissue Cleared 3D Brain Images

arXiv

引用

arXiv 2025年

作者： Nazib, Abdullah Fookes, Clinton Perrin, Dimitri School of Computer Science Queensland University of Technology BrisbaneQLD Australia School of Electrical Engineering and Robotics Queensland University of Technology BrisbaneQLD Australia Centre for Data Science Queensland University of Technology BrisbaneQLD Australia

We developed a generative patch based 3D image registration model that can register very high resolution images obtained from a biochemical process name tissue clearing. Tissue clearing process removes lipids and fats from the tissue and make the tissue transparent. When cleared tissues are imaged with Light-sheet fluorescent microscopy, the resulting images give a clear window to the cellular activities and dynamics inside the *** the images obtained are very rich with cellular information and hence their resolution is extremely high (eg .2560 × 2160 × 676). Analyzing images with such high resolution is a difficult task for any image analysis *** registration is a common step in image analysis pipeline when comparison between images are required. Traditional image registration methods fail to register images with such extant. In this paper we addressed this very high resolution image registration issue by proposing a patch-based generative network named InvGAN. Our proposed network can register very high resolution tissue cleared images. The tissue cleared dataset used in this paper are obtained from a tissue clearing protocol named CUBIC. We compared our method both with traditional and deep-learning based registration *** different versions of CUBIC dataset are used, representing two different resolutions 25% and 100% respectively. Experiments on two different resolutions clearly show the impact of resolution on the registration quality. At 25% resolution, our method achieves comparable registration accuracy with very short time (7 minutes approximately). At 100% resolution, most of the traditional registration methods fail except Elastix registration *** takes 28 hours to register where proposed InvGAN takes only 10 minutes. © 2025, CC BY.

关键词： Image registration

来源：评论

学校读者我要写书评

暂无评论

Molecular architecture of mammalian pyruvate dehydrogenase complex

引用

Protein & Cell 2025年第1期16卷 72-78页

作者： Maofei Chen Yutong Song Sensen Zhang Yitang Zhang Xudong Chen Minghui Zhang Meng Han Xin Gao Sai Li Maojun Yang Tsinghua-Peking Center for Life Sciences Beijing Frontier Research Center for Biological StructureSchool of Life SciencesTsinghua UniversityBeijing 100084China Ministry of Education Key Laboratory of Protein Science Tsinghua UniversityBeijing 100084China State Key Laboratory of Membrane Biology Tsinghua UniversityBeijing 100084China Department of Biochemistry and Molecular Biology School of Basic Medical SciencesShenzhen University Health Science CenterShenzhen 518060China Protein Research Technology Center Protein Chemistry and Omics PlatformSchool of Life SciencesTsinghua UniversityBeijing 100084China Computer Science Program ComputerElectrical and Mathematical Sciences and Engineering DivisionKing Abdullah University of Science and Technology(KAUST)Thuwal 23955-6900Kingdom of Saudi Arabia Computational Bioscience Research Center King Abdullah University of Science and TechnologyThuwal 23955-6900Kingdom of Saudi Arabia Cryo-EM Facility Center Southern University of Science and TechnologyShenzhen 518055China Beijing Life Science Academy Beijing 102209China

Dear Editor,Pyruvate dehydrogenase complex(PDHc) is a large multienzyme assembly(Mr = 4–10 million Daltons) consisting of three essential components: pyruvate dehydrogenase(E1p), dihydrolipoyl transacetylase(E2p), and dihydrolipoyl dehydrogenase(E3). These three enzymes perform distinct functions sequentially to catalyze the oxidative decarboxylation of pyruvate with formation of nicotinamide adenine dinucleotide(NADH) and acetyl-coenzyme A(Patel and Roche, 1990).

关键词： sequentially oxidative Dalton

来源：评论

学校读者我要写书评

暂无评论

Grid Impedance Adaptive VSG Control Based on Accurate Small-Signal Modelling

引用

IEEE Transactions on Industry Applications 2025年

作者： Wang, Minyang Yang, Yuxin Zhang, Yuchen Ambia, Mir Nahidual Ahmed, Al-Durra Zhang, Daming Meng, Ke Zhang, Ge University of New South Wales School of Elec Eng and Telco Faculty of Engineering NSW Australia Queensland University of Technology School of Electrical Engineering and Robotics Brisbane Australia University of Sydney Sydney Australia Principal Grid Connection Engineer at Nordex Oceania Australia Khalifa University Electrical Engineering and Computer Science Department Abu Dhabi United Arab Emirates ElectraNet Pty Ltd Adelaide Australia University of Queensland School of Electrical Engineering and Computer Science Brisbane Australia

The wide-spread integration of renewable energy has increased inverter-based generation, weakening grid strength and inertia, bringing challenges to power system stability. To address these issues, virtual synchronous generator (VSG) control has emerged as a modern grid-forming (GFM) technology that can autonomously provide active and reactive power to support grid frequency and voltage. However, when a VSG inverter operates under strong grid conditions, power oscillations may occur under a fixed parameter setting, making the grid difficult to reach a steady state. Under such circumstances, traditional stability analysis based on the rate of change of frequency (RoCoF) is insufficient, advanced analysis methods based on the changes in system impedance and further improvements are required. This paper proposed an accurate small-signal model of the VSG inverter along with an impedance identification method for grid impedance. Additionally, an improved VSG control adaptive to grid impedance is proposed based on the knowledge of small-signal model and system impedance. All proposed methods have been tested on a single-machine-infinite-bus (SMIB) system and the stability results demonstrate their effectiveness under both weak and strong grid scenarios. © 1972-2012 IEEE.

关键词： Synchronous generators

来源：评论

学校读者我要写书评

暂无评论

ACO-GA-Based Optimization to Enhance Global Path Planning for Autonomous Navigation in Grid Environments

引用

IEEE Transactions on Evolutionary Computation 2025年

作者： Heng, Han Rahiman, Wan Universiti Sains Malaysia Engineering Campus School of Electrical and Electronic Engineering Penang14300 Malaysia Universiti Sains Malaysia Engineering Campus Nibong Tebal Penang14300 Malaysia Daffodil Robotics Laboratory Department of Computer Science and Engineering Daffodil International University Dhaka1216 Bangladesh

Exploring path planning techniques for autonomous vehicles to find the safest and quickest routes is fascinating. Among these techniques, generating grid-based maps is notable for simplifying the path planning process, especially in complex environments. This study introduces a novel optimization approach, the modified ant colony optimization and genetic algorithm (MACOGA), which is tailored for path planning in grid environments. MACOGA combines ant colony optimization (ACO) with genetic algorithm (GA) to efficiently navigate grid spaces. First, ACO identifies potential routes within a grid map, and GA optimizes them to find the optimal route. A key feature of MACOGA is its probabilistic prediction mechanism, which improves node selection by integrating the heuristic factors and probabilistic calculations of feasible potential paths, which significantly increases the likelihood of generating feasible initial paths and attains higher success rates in various environmental tests. The enhanced GA in the MACOGA incorporates ACO's pheromones and heuristic factors, and introduces adaptive probability crossover and mutation operations. Additionally, a path reconnection operation ensures favorable length and smoothness. To evaluate MACOGA's effectiveness, six planar models of varying sizes and complexities were studied. Results show MACOGA's superiority in path length and execution time, achieving the shortest path and planning time in several environments. The shortest path can be optimized to yield an improvement of up to 6%. Regarding runtime, even in the complex scenarios, it executes in just 0.3 seconds, outperforming other comparative algorithms. Furthermore, the success rate can reach 100% in relatively complex environments. © 2025 IEEE.

关键词： Ant colony optimization

来源：评论

学校读者我要写书评

暂无评论

Multi-Keypoint Affordance Representation for Functional Dexterous Grasping

arXiv

引用

arXiv 2025年

作者： Yang, Fan Luo, Dongsheng Chen, Wenrui Lin, Jiacheng Cai, Junjie Yang, Kailun Li, Zhiyong Wang, Yaonan School of Robotics Hunan University China National Engineering Research Center of Robot Visual Perception and Control Technology Hunan University China College of Computer Science and Electronic Engineering Hunan University Changsha China

Functional dexterous grasping requires precise hand-object interaction, going beyond simple gripping. Existing affordance-based methods primarily predict coarse interaction regions and cannot directly constrain the grasping posture, leading to a disconnection between visual perception and manipulation. To address this issue, we propose a multi-keypoint affordance representation for functional dexterous grasping, which directly encodes task-driven grasp configurations by localizing functional contact points. Our method introduces Contact-guided Multi-Keypoint Affordance (CMKA), leveraging human grasping experience images for weak supervision combined with Large Vision Models for fine affordance feature extraction, achieving generalization while avoiding manual keypoint annotations. Additionally, we present a Keypoint-based Grasp matrix Transformation (KGT) method, ensuring spatial consistency between hand keypoints and object contact points, thus providing a direct link between visual perception and dexterous grasping actions. Experiments on public real-world FAH datasets, IsaacGym simulation, and challenging robotic tasks demonstrate that our method significantly improves affordance localization accuracy, grasp consistency, and generalization to unseen tools and tasks, bridging the gap between visual affordance learning and dexterous robotic manipulation. The source code and demo videos will be publicly available at https://***/PopeyePxx/MKA. Copyright © 2025, The Authors. All rights reserved.

关键词： Vision

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：